Attendance:

• Cary Phillips
• Christina Tempelaar-Leitz
• John Mertic
• Joseph Goldstone
• Kimball Thurston
• Larry Gritz
• Nick Porcino
• Peter Hillman
• Rod Bogart

Guests:

• Sergio Rojas

Key Dates to Review:

• Tuesday June 1, 2021, Imath v3.0.4 release
• Thursday June 3, 2021, OpenEXR v3.0.4 release
• Sunday June 20, 2021, Imath v3.1 feature freeze
• Sunday June 20, 2021, OpenEXR v3.1 feature freeze
• Monday June 28, 2021, Imath v3.1.0 code freeze
• Wednesday June 30, 2021, Imath v3.1.0-beta release
• Friday July 2, 2021, OpenEXR v3.1-beta code freeze
• Sunday July 4, 2021, OpenEXR v3.1.0-beta release
• Monday July 26, 2021, Imath v3.1.0 code freeze
• Wednesday July 28, 2021 Imath v3.1.0 release
• Friday July 30, 2021, OpenEXR v3.1.0 code freeze
• Sunday August 1, 2021, OpenEXR v3.1.0 release
• August 4-5, 2021, ASWF Open Source Days

Agenda:

• Confluence space
• Key dates
• Release process and terminology
• Outstanding PR's
• v3.1 Release progress
• Recent bugs

Discussion:

• Bazel is broken, the version number wasn't updated, so a bazel build of v3.0.4 will say it's 3.0.2.
• Is the libtool-style so versioning necessary?
  • Kimball: Yes, it's standard, projects should use it and many don’t. We should keep it.
  • SOVERSION is more typical c library approach.
  • Larry: Why isn’t semantic versioning OK?
  • https://www.gnu.org/software/libtool/manual/html_node/Updating-version-info.html#Updating-version-info
  • Kimball: would be nice to be able to say our names are always major.minor. The loader is only ever looking for SONAME.
• The new Confluence space holds the TSC notes, so no more cluttering up the repo and PR's.
  • The ASWF OpenEXR calendar has key dates. for the release schedule.
• Kimball: Imath “config” target is for headers only, no library
• 3.1 release:
  • Kimball: need more tests
  • Documentation should make clear that 3.1 is additions, don’t have to rewrite everything.
  • We need a Roadmap confluence page that summarizes where we're going with this, what to expect in 3.2 and beyond.
• PyImath changes from ILM, to bring the two into alignment.
• IMATH_NOEXCEPT should be a CMake option.